SPEAKER_00 0:00

You are listening to the AI Investor Podcast from 24-7 Wall Street. On today's episode, we look at Micron's earnings, Google's impressive bounce back thanks to the three most important letters in AI right now, and that's RSI. We look at an update on Vicor, and then we do a quick QA and stock updates. All that and more is next. Eric, great to see you. Uh how are things in the AI world today? I've I feel like the my portfolio has been very uneven, right? The software trade is waning, the chips trade was back, and then the chips trade gave a little back. You know, what's your take on the market right now?

SPEAKER_01 0:35

Yeah, from a macro perspective, it continues to get bleaker. We still have oil at significantly higher prices. Tenure treasuries now hit their highest level since 2002. Um, and the situation around high rates has really gone worldwide. Japan's at its highest level in three decades. Uh, spread to Europe, Germany is at its highest level since 2008. So we're seeing all this, but against this backdrop, we do continue having positive news in AI. We'll talk about Micron's earnings, like you said, that's rapidly become kind of the Super Bowl behind NVIDIA each earning season, just kind of the numbers they're putting up.

SPEAKER_00 1:11

It's the puppy bowl to NVIDIA's Super Bowl, which is still a big deal and still gets still gets millions of viewers, right? More than most sporting events on Earth.

SPEAKER_01 1:19

Well, it's interesting. Uh they're kind of like a bad Super Bowl at this point, too. I'm looking at Micron, some really outstanding numbers and just kind of a yawn down 1.6%. But, you know, that's what happens when the market, um, the consensus is set for greatness. So we'll talk about that, but Austin, we'll we'll also catch up. We've had one of the stocks that I had featured in that large presentation that we uh did a few weeks ago has seen significant price rises. And we wanted to get to some QA as well. So we'll we'll kind of go through the mailbag.

SPEAKER_00 1:50

Uh wonderful. Eric, before we jump into that, I do want to say one more thing. You have a lot of stocks in the portfolio, dozens of positions. We had mentioned this on a prior pod, and although we do not take sponsorship and we've never done advertising on the podcast, I do want to shout out to our partner at Yahoo Finance for their product Alpha Space. Uh, you and I have actually tracked the AI portfolio in Alpha Space, and it makes it really easy to track earnings across dozens of positions, see news for dozens of companies all at once. I know you've made the Google Doc available for people to track your buy recommendations and the current price. But if you really want news across the whole portfolio, and particularly for during earning season, uh, we do recommend our listeners head over to Alpha Space and sign up for a free trial. We do have a link in the show notes that you're able to uh access that product and we will get a small compensation for it. But really, it's just a great product that we recommend. You and I have used Yahoo Finance for years, and I think this is this is the best iteration they've ever had of a portfolio tracking tool, which is great for something like the AI investor portfolio, where you have dozens of positions. Um, you had talked about Micron's earnings. So let's talk about that. We're talking about earnings, strong performance, stock flat. I'm having a really difficult time squaring how much of this is company specific, industry specific, or is it macro headwinds, right? So Micron itself is putting up strong performance, but that was also expected. So maybe we could say that's baked in. We could also say it's just waning enthusiasm for the memory or the AI trade broadly, or it could be any of those other things you listed at the beginning. The 10-year treasury is at the highest level since 2002. Uh, we've got oil still perpetually high, inflation seems to be untamable. So, what is your read on this muted reaction in the AI industry we've seen and micron in particular uh the last few weeks?

SPEAKER_01 3:35

Yeah, as I alluded to earlier, it's it's largely a stock that greatness is expected. Uh the this report was great, revenue up 379% from last year, uh, which is such a cartoonishly large number that we've gone used to. Uh, guidance for next core is 61.5 billion, which beat. But Austin, as we saw from the peak of NVIDIA, the beats start getting smaller, right? And and you know, that's kind of the market is expecting such a huge number. And there's these whisper numbers from the buy side, and it becomes harder and harder to be able to give those beats, even while overall performance continues to be great. And adjusted gross margins, um, they're actually projected to decline next quarter down to 86.25. We've talked about margins, they won't be able to stay at these levels forever. We all know this is happening. The question is, how enduring, once this cycle doesn't have these extreme dynamics, will kind of the memory market be able to stay at? So let's talk about a few things from these earnings I found interesting. First, we had a quote from the company we expect memory and storage supplied demand conditions to be much tighter in calendar 2027 and 2028 than they were in 2026. And they continued. How is that even possible?

SPEAKER_00 4:51

How is that even possible? I just I mean, you said we're we're dealing with cartoonishly large numbers. Yeah. But also, I mean, that's a product of cartoonishly large demand, cartoonishly large capex cycles. How is it possible that things are tighter than they were the last few years? And isn't that the isn't that the counterpoint to the bears who are saying, you know, this is a cyclical stock that's about to crater? And let's be clear, right? We know there's gonna be, you know, supply clutch and supply shortages, the way this industry works, but we are in a fundamentally different cycle. It does not follow the the prior patterns. Um correct.

SPEAKER_01 5:22

Yeah, and one of the big factors that we're seeing is memory post uh you know early June when we saw kind of excitement AI stocks really hit their peak. Um it had gone down substantially, both micron and stocks like Sandisk. And then Sandisk releasing its guidance, as we had covered in the show, that they expected adjusted margins to stay at 80% through 2030, is what really led to a rebound in a lot of these stocks. So memory stocks, they are off their peak. Um, but you know, obviously, if you look at a one-year number or you look at a year-to-date number, it's still quite outstanding. And compared to some other segments of the AI trade, where these stocks are relative to how close they are trained to their recent highs, their 52-week highs, it's actually held in better than other segments. And and Micron did add to this story today because what they hadn't talked about before, they had talked about 2027, they added 2028 today. So this continues pushing the story ahead. And and as we're talking about how long this kind of elevated environment goes for, well, they talked about 75% of their 2027 output being covered by these agreements that they're signed with customers. But they also said 35% of revenue through 2030. And another thing, well, gross margins will trough next quarter. Well, there's a few unique conditions likely causing that. And they said they actually expect margins to then grow throughout fiscal 2027, which would be above consensus. And that's largely driven by improvements in high bandwidth memory. And Austin, the last point here is they talked about beginning to do large share buybacks. So if you're looking, if you're a micron investor, which we did invest in micron, I believe for $83 a share or something very low at the beginning of 2025. Um one one of the you know key catalysts that you could have right now is number one, continuing to show a longer and longer duration for kind of these current environments, continue to show elements that can keep margins elevated beyond kind of these historical norms such as customization in memory and and long-term dynamics within AI. And then the third thing is what you're going to do with this capital, because there's a good chance in the next two years, Micron will make, in terms of cash flow, something like a third of its current market cap. So if they continue to invest this back into the stock and show that they're going to use this capital aggressively on buybacks, that becomes a relatively near-term catalyst. So I think in terms of positioning for Micron, you know, it even with the outstanding returns over the past year, year to date, we continue to hold this stock and we continue to like what this report says about the broader memory space, which again, beyond beyond recommendations like Micron and SK Heinrichs we had previously issued, we also have a lot in the kind of companies selling to these companies for future capital expenditure. So I liked what was said about that in this report as well, with some um commentary on CapEx coming ahead of expectations next year.

SPEAKER_00 8:40

Can we talk a little bit about buybacks for a moment? So it it is an extremely positive sign that Micron is announcing these buybacks. They feel good about their share price. They're going to be puking so much cash, as you said, that it's inevitable. This also comes right on the heels of NVIDIA announcing a $150 billion increase to their share buyback program, which I think takes their remaining capacity up to like $240 billion. Again, speaking of comically large numbers, their buyback is now larger than some major percent of companies in the SP 500. Like that's outrageous on its own. But I also am stuck in a little bit of a moment of tension here. On average, I think the data has shown that share buybacks from companies are generally poorly timed as a blanket strategy. We take all the buybacks that are done, they are generally uh capital destroying. However, there are a few exceptional cases in history, um, many actually, where buybacks are one of the major factors that goes on a decade-long tail run for very consistently profitable companies. Apple's probably the best example here. Apple is one of the best performing stocks of basically any time frame. And a very large percent of that is how consistently they have been making cash and buying back shares and the amount that they've reduced their shares since they've begun that share repurchase program. So if we look at a micron, we look at an NVIDIA, where do you sort of square those? Are we seeing companies that are temporarily extremely flush with cash and probably poorly timed on the buybacks? Or is this more like an Apple cycle where we have fun companies that are in a fundamentally different paradigm where they're going to be high margin, high cash flow creative for like a decade, and therefore things like these buybacks make a lot of sense? Where do you draw that line?

SPEAKER_01 10:20

Yeah, the the technology cycle in buybacks is actually really interesting because again, you had these companies in a similar position. They're making so much money, they just need to do something with it. You know, you can't, you can't be investors are gonna start wringing their hands if there's no capital returns and you're placing $300 billion on your balance sheet. The alternative to that though, uh, you know, would would Apple like a little bit more cash that it could maybe make some more investments into AI today? Would would Google have appreciated not having some of those buybacks that it would not be need to tap the debt markets or it would have more flexibility? So a lot of these companies that had issued buybacks um, where technology ended up going, that that capital, they would actually really like to have that capital right now. So I do think over time Apple's was positive. It did change the stock around. And and sometimes that's too, you know, if if the market doesn't believe you're going to be able to invest that money, you need to show a willingness um, you know, to be able to return it. So I think that's really the position Micron's in. They will have some ability to invest, but you know, also if they invest too deeply, um, it could be something that would impact their gross margins, which the market also wouldn't like, right? So you you have those dynamics. And with NVIDIA, their alternative has largely been uh creating these funding mechanisms in the market, but that's created its own questions about kind of this um kind of the self-financing issue, right? So I've personally liked what I've seen from both of these companies. I would rather a company like a micron, um, I would rather that they focused on buybacks right now than maybe doing something like hiking their dividend dramatically. Because once you hike your dividend dramatically, when a cycle changes, you're gonna have to dramatically cut your dividend, which is gonna create its own kind of negative sentiment cycle for the stock. So as of right now, I think they're just in an incredibly unique situation. It's it's it's that it's essentially become the companies that used to get all of the cash flow are now pouring it into a slim group of companies. The Googles that used to be the cash flow, uh kind of just piling it in a Scrooge McDuck. They're swimming in their pool of cash. You know, that that pool of cash is getting directly funneled into Micron and NVIDIA. And so the the problem's almost been transferred. And and I use problem, Austin, as loosely as possible. They are making so much money.

SPEAKER_00 12:48

Some problems are better than others. It's like it's the Walter White problem, right? Like, where do I hide all of this money? Um so okay, so I a great point. I agree with all of that. One of the things you had talked about there is that tech companies, one of the reasons that they want to retain cash or they should want to retain cash, one of the lessons from history, is when you do hit these massive RD cycles, you want to be there in a great position to take advantage of it, to invest well, to make products for the next generation. And those are kind of hard to anticipate. You talked about Google and Apple there. One of those areas of RD and major investment for Micron is physical AI, right? The transition to robotics, um, the ongoing needs of self-driving vehicles. So talk to me a little bit about Micron in that space and their comments about physical AI. And while some of their cash is going to be going to buybacks, a lot more also could be going to these internal RD projects.

SPEAKER_01 13:37

Yeah. And as we think about where robotics is, how how long until it scales, what the future looks like, NVIDIA's long been a really interesting company to follow. You want to look at on the transcripts, what what um they're they're saying about robotics, what they're saying at conferences. Micron's increasingly becoming another company to follow because they are now starting to give quarterly updates on the robotics market and and what that could be mean as a next driver for memory. So on their call, they said level four autonomous vehicles are going to use about 200 gigabytes of memory, which is a lot, uh, multiple terabytes of storage. That's roughly 10x what most of the self-driving cars on the road today, because right, these aren't full featured self-driving cars. These are what you call a level two or a level three. So they're seeing a 10x increase, but they expect, and this is what's big, I think, humanoids to have comparable memory and storage needs. So as they said in direct quote, physical AI can become a significant driver of memory and storage demand by the end of the decade, and that they have customers already sampling. So, Austin, on one hand, it really behooves these companies to always be talking about what's next, right? In technology, you're never standing still. There's always concerns about concentration towards one trend, like AI data centers. They want to talk about catalysts for what's going to drive their business beyond where they're concentrated today. On the other hand, these are the companies that often have the best line of sight into what's happening in the space. So the fact that Micron already is engaging with customers on next-gen products for robotics is something to watch. And it's just going to depend how much scale we have, right? This the car market, people got excited about semiconductor companies moving into it in 2010. And and back when I was following, that was a really big kind of demand calus for NVIDIA you'd read about. But you always need to go and actually sharpen your pencils and do a spreadsheet and say, ah, this market's just going to be limited fundamentally because how many cars you'll have. In in a self-driving world where where the majority of cars are self-driving, you know, you can you can get to some impressive numbers. But if humanoid robotics and robotics truly takes off in a significant way, we have breakthroughs, et cetera, you could be looking at something like billions of units. And if memory needs were that high across it, well, that that would create a significant demand cycle that that would play into memory much the same way AI and LLMs have played into it. So I think it's something to watch. I think whenever I'm watching Micron's earnings calls now, I'll be looking at the numbers, I'll be looking at the commentary about memory. And then a third factor, I'll be looking at what they say about robotics, which um, you know, it's probably not what anyone expected a year ago to be looking at micron for the future of robots.

SPEAKER_00 16:36

Well, the the future is stranger than what is it? Uh reality is stranger than fiction, right? And yet here we are. Close enough. So close enough, close enough. Um let's talk about uh Gemini. I want to hear about the Argon model release. And this is, you know, I'm not sure if all these model releases cluster around each other on purpose, or that's just the way the training cycles have gone. But you know, we we've had ask we've had OpenAI rolling out their dots agentic model, which is sort of an answer to Meta's Muse, which recently rolled out. Now we've got Gemini for Argon. I've only had time to play with one. I will say Meta's Muse is extremely impressive, um, remarkable. I've not played with Dots yet. I'm excited to play more with Argon, but what are you seeing in this model? And does Google have something to be proud of here? We had talked in the last episode about how we fully expected Google to quick follow with a agentic model. Is this it? Are we still waiting?

SPEAKER_01 17:30

Well, across the entire summer, as we've talked about on the show, coding has become a huge differentiator in capability and in building models and also how you get revenue from your models. So Google had fallen behind in that zone. And across the entire summer, there was talk of releasing a new model that would put them back on the frontier, competitive with the other companies like OpenAI and Anthropic, and it was delayed time and time again. So we're finally seeing it here. And Austin, the real goal is closing the gap on coding, enterprise work, cybersecurity, a lot of places where the money is being made in um kind of the broader LLM space. So they released this, and and you know, we could put the benchmarks on screen. Um, you know, there's a lot of areas that uh they're showing they are competitive or beating uh the top models from both anthropic and open AI. I'll give a couple caveats though. You know, it seems to be the weakest on coding still. Um, the other point I'll get into shortly is that we are entering a new world where the best models from anthropic and open AI are often not released to the public. Well, the companies that are chasing them have more incentive to get their best models out quickly. So we'll need to keep that in mind. Now, a few a few things I like about Austin and play um kind of into uh you know what's happening in this battle between all these companies. Number one, it's 39% cheaper than Astra from OpenAI, 67% cheaper than Opus. So it's it's working to kind of undercut them on pricing. Second, we've talked about this world where the next kind of step change in agenc capability is these longer tasks. And we see this has a 1 million output token limit. Opus uh 5.5 from Anthropics 128K. The previous Google Max was 64k tokens. So this just means these long-running tasks where more is happening in each cycle are getting more and more capable. And uh, I I guess I'll just say, you know, at at the end of the day, you had you had teed up kind of the three most important uh letters in AI right now are RSI. I I think it's we've seen not so subtle hints from people at DeepMind, which is you know, creating Google's uh most cutting-edge AI models, that they they've made some breakthroughs there and they they believe that this is going to be a differentiator for them. And I I also saw some quotes from Google engineers that they're going to continue trying to improve Gemini 4 on almost a weekly basis. So you look at what RSI could do, it it would improve both the quality of the models, but also the speed of releases as well. So you're seeing some people from the company hinting that that's going to be kind of a next uh kind of iteration from Google. So at the end of the day, just to kind of close up this this specific segment, you had talked about how it feels like we're getting this rapid fire release from all these companies that felt like anthropic and open AI were in the lead. And suddenly it feels like Meta and Google are closing the gap, and we've got XAI and Elon Musk lurking in the background. So, so so what's going on? What does this mean? I I would say number one, what we probably have is, you know, sometime late last year, early this year, op uh, I should say anthropic, they started really increasing the cadence of their development. And they they got a large lead. And this was in the wake of kind of the cloud code breakthrough. I think open AI, whether that was taking down some safety guardrails or whatever they needed to do, they worked really hard and they've caught up with Astra. And now they're also catching up on a revenue side. But both of these companies now they're leading models, as I said. They're they're often not released to the public. They they keep them internally and they're releasing capabilities that that are behind what they have uh with within the company itself. The companies that are chasing have every incentive to get their newest models out faster. So I don't know if the gap of uh Google or a Meta being six or nine months behind has has really closed. Or if it's just the fact that the dynamics have changed that OpenAI anthropic, you know, they are not necessarily releasing their fastest models while other companies are. And second, I'll say at some point we're going to see a shift, I think, from the well um relative quality of foundation models and these kind of benchmarks you see whenever they all come out, towards the quality of distribution business models. And with that, we see open AI anthropic with large established leads in enterprise that both Meta and Google, Meta hired MongoDB's CEO this week to run an anthro uh uh enterprise division. Google is is obviously they're trying to undercut these companies with the pricing here. But both these companies then have the advantage in distribution. So Meta and Google can get to consumers, and that's where that's where we see something like you mentioned OpenAI's dots. That's trying to fight back against this distribution advantage for Meta. And then somewhere in the background, we just have XAI as a chaos agent. Elon Musk is just lobbing things out there. He's he's trying to disrupt the market. So I think as we look at the the end of the year, we're moving into this phase where models increasingly improve themselves. We are getting to a stage where I think we're gonna stop focusing so much on any individual model, and we're gonna focus more on kind of the strategy and which niche these companies have. And I think the market will increasingly focus on that as well.

SPEAKER_00 23:33

Can we just take a moment and appreciate what a remarkable time this is to be a consumer and the amount of consumer surplus that is coming out of this? And that's a concept in economics, right? Which is just like the amount that societies benefit or people benefit from advancements like this. You have some of the wealthiest, most capable, most impressive companies on earth pummeling all of their money into technology and putting pressure to make it cheaper and better and faster. I mean, you talked about the token increases on Google's model. You talked about the pricing and cost efficiencies. That that might hurt Google's revenue, but it is fantastic for the rest of the world, right? It's fantastic for small businesses that want to deploy AI. It's fantastic for users who want to deploy AI. We are such beneficiaries at this moment, and I just I want the amount of competition in this space right now is so fierce and so deep pocketed, and we are the beneficiaries of it. I meta's Muse model being, you know, or tool rather being available for free and as capable as is remarkable. It's incredible. I'm I'm I the this level of competition, the amount that we as consumers and businesses are benefiting is incredible. I'm excited to see where this goes. Enough of the stump speech, though. Um, Eric, let's get on to one. I'm not gonna call this one that got away, but I think this is gonna live next to Bloom Energy and the one that's gonna bother you and bother our listeners a little bit because you did in fact recommend it. Maybe you own it personally, I'm not sure, but it never made it into the AI investor scorecard. What's going on? What's going on with this one that semi-got away? Well, we we you know, I don't want to get too hung up.

SPEAKER_01 25:08

I don't want to get too hung up. You know, I made it, I made it one of I did three um, whether you want to call them buy recommendations or ideas in the Investicon speech I done. And and then we put that into the 9-11 podcast, the one we released on September 11th with a long presentation on AI. I did a breakout on three different stocks. One was uh Semtech, one was Vicor, and one was FTAI. And I I wanted to catch up on Vicor a little bit because what what happened with Vicor uh since it's become one of the fastest moving stocks in the market, almost immediately following when we recommended it. Uh it's up more than 50% in in less than a month. So let's quickly talk about what's happening with the stock. There might be some people who missed that. Um, you know, it was part of an almost two-hour podcast in the middle of it. It was on our YouTube channel as well. Um, I hope people went out and and you know looked at it as an idea because it is exciting. Um, but basically Vicor, which is ticker symbol V-I-C-R, it's the next evolution of our power management theme we've talked about throughout this year. In the past, we talked about the data signers itself. How is how's the architecture with uh this 800 V transition? How is the data signer going to change? Vicor is all about how it changes delivering power to the chip itself. So they've got key technology known as vertical power delivery. It lowers voltage close to the chip. Basically, this shortens the high current pathway. And how I described this in the presentation uh that that we had given on that September 11th podcast was you think about high bandwidth memory. Really, this is all about shortening a pathway to the actual die to get more bandwidth. You think about co-package optics, it's a similar idea. This is shortening pathways that cause some kind of performance degradation. It's using this advanced packaging of chips to get around limitations. And we have something similar happening right now in how we deliver power to the chips. So Vicor has a couple different paths of how they can make money. They can sell chips or they can license their technology. On their last earnings, which was September 21st, they they announced a new licensing agreement. Um, third-party modules covered by their pants uh would allow them to raise revenue outlook by 20% sequentially. So that's that's into their next quarter. And last night they raised um they raised that to more than 30% sequentially. So we see the stock up another 10%. So also I just wanted to cover this and we'll get into QA in a minute because this was a stock we were excited about. I know a lot of people have asked for our thoughts on power management. We'll cover this a little bit more in the QA. But this was the most recent idea we had given and uh it's moving really quickly. And and the problem is this is gonna be probably a pretty asymmetric stock, I believe. Either, either there's gonna be uh wide uptake of vertical power delivery, um, or it's gonna be a stock that proves to be relatively expensive where it is today. But I like having some of these in the portfolio that there's a potential almost architectural change in technology and you can get on the right side of it. I I always want to have some bets like that in the portfolio. And and to date, it seems like it's working pretty well, right? We we've seen this this extreme takeoff in it. So uh congrats to any investors who bought it. And uh for anyone who it was buried an hour into a presentation, you didn't get it. Uh, well, I I'm sorry you missed it, but I don't think it's too late to catch up on the story.

SPEAKER_00 28:55

Uh wonderful. And congratulations to anybody who picked it up at InvestiCon, I believe. And that is a great transition to our next question. Uh, and again, this is about the community. We like to hear from our fans, so I'm glad we're doing another QA episode here. This question from Bob Dylan 1409 via Spotify. Brilliant, Eric. I was at your talk at Dublin, speaking of InvestiCon. I voted for you for best speaker. Oh, that's nice. Thank you. Uh, I was just wondering what you thought of Nebbius, uh Nebius group. Uh, would you put them in the same category? And this says, I'm not sure what the same. Okay, so Oh, yeah, yeah, yeah.

SPEAKER_01 29:28

This was just uh, I put a note to you that I I don't know what the category he was referring to is. Um, so I I can't speak to that. But we have talked about Nebbius. If if you're new, if you discovered the podcast after that Dublin speech, we have talked about it in some recent podcasts. The the kind of 10,000-foot view is from a business model perspective, I really appreciate the most how the Neo Clouds it started its business trying to differentiate on software that that puts it really in a better competitive standing, especially as most of these companies go for scale. And and we have, in terms of capabilities and size and scale, Nebus and Core Weave really stand out from a lot of the competitions. But, you know, we it the NeoCloud space is is hard to evaluate because you you have that, and then you have how quickly companies can stand up projects, who who has energy availability for their projects. And those are areas where a lot of people are drawn to stocks like SpaceX or IRN. So I think you know, this space is worthy of a deep dive because it requires so much attention to detail. And uh I think we'll we'll add that and we'll try to get into that in Q4, but especially with where financing is at, um it that's going to have the biggest implications broadly on the neo cloud market. So it's not something I I haven't added, um I haven't added Nebius to date. I've said it it's one that I wish, you know, when when we had first previewed Neo Clouds last year, I had said it, I thought it was the best opera in the space. I wish we had added with this performance. But um, you know, so much of investing, and I'll I'll cover this in a few questions, is trying to get exposure and trying to get what you believe is the best in breed across ideas. And uh we we tried covering the compute space by getting a stock like Amazon, which we said could benefit. I can see Amazon Web Services being a trillion dollar uh revenue uh business over time, but we wanted to eliminate some of the risk. And and you know, we had the thesis that over time these larger hyperscalers would would remain very dominant and make it hard for these new entrants. Um, you know, so far the market is pricing it so that uh it's very dismissive of a company like Oracle trying to break into that, but the smaller hype, uh the smaller not hyperscalers, the smaller neo clouds um as a group have done pretty well. So so that's a little bit of a miss, and and it might be something that we need to evaluate as kind of a missing space from the portfolio.

SPEAKER_00 32:05

You had mentioned Oracle there. I I mean, isn't a big part of their overhang also just concerns about financing? I think they were they went pretty heavy into the dev space as well. It's it's less, I mean, there certainly could be some discounting on their value as a neo cloud, but I I think a lot of that is just financing concerns.

SPEAKER_01 32:22

Yeah, and that's where it's so hard. You know, you need to, it's not just that you can evaluate a company's technology. It again, it is what does a pipeline look like? What's financing look like? What's the quality of um the people leasing your projects? So it's it's not something you can wade into. Um, it it requires extremely deep research, I believe, to invest in this space well.

SPEAKER_00 32:46

Um okay, let's move on to the next one. We've got James Patrick, also from Spotify, saying, Great podcast, guys. We'll watch on YouTube as well. Thank you, James. And to all of our listeners, we do appreciate you watching on YouTube and commenting. It is uh the best place for us to get questions from the community. So thank you, James, for hopping over there. James says, looking forward to hearing about power companies to power all of the AI and grid infrastructure. Will you mention AIPO ETF? Okay. This is a big question because we have talked a lot about the powering needs. And, you know, I'm gonna go get a coffee and come back in 20 minutes when you're done.

SPEAKER_01 33:19

Yeah, I there's so many ETFs. Um, it it is hard to keep track of them. So AIPO, I believe this was from Defiance, which is a company that does a large number of uh, yeah, it is, a large number of ETF uh products. It looks like it's got about a billion dollars in assets under management, which is a fair size. You know, you look at this ETF, it really tells the story of AI this year. The last three months, it's down 15% as kind of interest in the AI space has waned from an investing perspective. Year to date, it's up 26%, which is outstanding. You'll always take a 26% year-to-date gain, especially in a year that so much from a macro perspective has not gone the way that we would have hoped for. But the ETF itself, it seems reasonable. I'll always look at the top holdings. Um, Eden, Quanta, G, Vernova, Verdev, Bloom Energy, those are all stocks we've spoken warmly of. Uh, if you listened to the interview we had with John Rotanti, there will be a very high overlap with the kinds of industrial companies he has in his coverage zone. I did know some scope creep right after those companies, which fit into a, you know, a pretty tight association than its Alphabet NVIDIA Broadcom, which, you know, it just feels a little bit out of place. I don't know if you need those in a specialized ETF, but um, overall, I think this ETF looks strong. My biggest nitpick would be it's all US listed stocks. And one of the biggest advantages of ETFs is you can use them to get exposure to parts of the world that can be complicated to buy on your own. You know, we we try keeping international recommendations down. We we often do them as recommendations without, you know, putting them in the portfolio because we know how hard it is uh for many investors to go out and buy a stock from Taiwan or many of these other exchanges. But ETFs can do that for you. And that's where a space like robotics having an international ETF, it it meaningfully makes investing in the space much easier. So I I wish they had done that because they're they're missing on some areas. If anyone out there um knows of some ETFs, because there are so many ETFs now that cover this power management theme and have more of a global purview, uh, we'll we'll try and mention that in a next episode, because that would be the biggest area for improvement I see.

SPEAKER_00 35:44

And if anybody has not watched your ETF explainer video on YouTube, I would encourage them to hop over and check it out. It is fantastic and characteristically well produced. Also, one of the things that you have talked about a lot in this podcast is you know the trend first and then the companies. And that is one of the nice things about the basket of ETFs. If you're not an expert in robotics and uh power management and plastics and um all of the things that go into actuators and robotics and GPS, if you're not an expert in those things, you can just buy a basket with an ETF and bet on the trend without um being an expert at your depth. Uh, another question here from Chad's organics from Spotify. I doubled down on on semi this week with the big drop. Curious on your thoughts.

SPEAKER_01 36:27

Look at look at that transition there. Going from investing in plastics to uh Chad's organics. Plastics, they're gonna be big. So uh power management. Uh as I as I allude to earlier, Vicor was the first new idea we've we've introduced from the space, but it's been a consistent theme across the year. We had first introduced it on the March 11th uh podcast, which was an exceptional time to buy. But then we tried adding a couple of stocks we had introduced as formal recommendations in May that we had missed when we had first uh kind of discussed the theme. And that was very poor timing because that was at the point that anything speculative across the AI space was going to see um abysmal returns over the coming months. Um, so overall on semiconductor, it looks like a terrible investment from summer and a good one if you had invested in winter or early spring. What's happened with the company itself? They they took a large hit after a querying company named Synaptics. Uh the acquisition was largely seen as a distraction. It pushed them more into Robux and AI, but Synaptics has pretty niche applications that, you know, it wasn't broadly seen as a needle mover by the market. Um, so from today's prices on semi, you know, I think it it's it should remain to be attractive, if especially if the data center growth that the company could could receive from the changes that are happening to how data centers are built uh come through in the next few years, which I still think, you know, the timing might be a little lumpy, but but this is a situation that's still on track to play out. The broader question is what is the ranking of some of these power management plays? So, Austin, earlier in the year we had bucket the portfolio by themes, and then we had talked about what we like best, what we might be concerned with, if if there was a sell, if we were going to sell anything. Maybe, maybe we need to go through that exercise again. We could add something like a power management to to kind of talk about which ones, which ones we like the most from today. Because yes, I think an on semi is a stock that looks attractive given the trends that has behind it at this moment. But you know, if you were only adding one or two, what would be the names we add? And and you know, that's that's something I would need to go through and do the analysis before, you know, saying from exactly this moment. But I think we'll add that to something we'll cover in the fourth quarter of the year.

SPEAKER_00 38:54

Uh wonderful. And let's power on to the last question here from Dan Peach, also from Spotify. Dan said, I have loved listening to the podcast over the last 18 months. Thank you very much, Dan. And really hope this doesn't become a macro podcast. Us two, us two. Nobody wants to be macro. We occasionally have to talk about it. Geopolitics, interest rates, you get it, but no, we don't like it either. It is not our cup of tea. So, Dan, we're right there with you. Uh, so Dan said, there's already so much macro content out there. I want to hear about the current companies in the portfolio. What happened with AMSC last earnings? Is the thesis intact, et cetera? Plus companies on the watch list. Are the robotics companies at better valuations now? Love your work. Just my two cents. Dan, fantastic comment. I really like this one. You know, I'm a team robotics, so I want to hear about that. Um, I think Regal Rexnort has had a very rough debut on the scorecard. Meanwhile, I've seen Ouster rebound quite a bit. I think I scooped up some in um 20% or so ago. So we've got some divergent stories in the robotics portion of the portfolio. Um, and that's probably exposure that will grow in the next few months, I would expect. But Eric, what would you say to Dan, who and no macro commentary. Whatever you do, don't say interest rates and don't say elections. I don't want to hear it.

SPEAKER_01 40:08

Uh American Superconductor, yeah, they they issued disappointing guidance. They missed on revenue, their EPS was light. Um, I you know, this is just something that is going to cause uh stocks in this broader AI space to fall. If if if you're forecasting, there's so many stocks blowing the doors off on guidance and what they're doing. We looked at what's going on with a micron earlier, that if you're one of the few ones to miss, um, you're going to get punished. Now, the longer term question. Um I think from a portfolio perspective, I I've talked about this in the past where I want a portfolio where like 70% of stocks, 60 to 70 are gaining and and you know, 30% are down, which might sound counterintuitive, but I want to make sure I'm making bets across themes that sometimes a theme won't work out. Um, but I I I do want it in the portfolio still, because there's always going to be hindsight in what worked and what didn't. But if you go all in, if you are all in on something like NeoClouds, as an example, we talked about those earlier, it it could work fabulously, but you're really increasing your risk profile that even if AI as a total concept works out, you could have a situation where, you know, you could be down 60 or 70% in a few years, depending on how this market moves, right? So if the broader concept behind kind of the portfolio we've built for the AI investor podcast, investing in AI, is you want to capture AI growth. I think you just need to be diversified across the potential futures we could experience. And with energy, you know, there's there's a couple key pathways right now. There's grid infrastructure, which is going to be tying into the grid, just what it says, and behind the meter, which is again that these companies often they're not going to be able to get these grid connections because the grid's not adding capacity fast enough. So they're going to be able to power their data centers increasingly with things like natural gas turbines. So we've been focusing a little bit more on behind the meter recently. We've had more recent recommendations in that space, more attention, because I think the space is going to have some big catalysts coming. Right now, the market's negative on a lot of these stocks because they haven't actually done the show me phase. But I think the show me phase where it moves from promise to really massive growth is really going to be something that happens and accelerates throughout 2027. The grid side tends to be a lot more lumpy. Um, you know, there's there's just so much that goes into whether or not these investments in the grid are going to happen. But I do believe that if we transition to an economy in the US that uh is really dependent on providing cheap intelligence at scale, it's going to be a trend that has to happen across the next decade. So uh what happened with American Semiconductor last quarter, it is a company that it is hard to read. If there might be something unique to it, but but my my core read from what I saw was it seemed like this is just this lumpiness that's going to play out with a lot of these companies exposed to grid infrastructure. Um, and I think for the long term, what I want to do is I want to have some ideas that are grid infrastructure focused. I want to have some that are more behind the meter. So is American Superconductor the best stock for this grid side? Maybe. You know, we've made it one of our recommendations, but I think we will continue adding even if there's been some recent weakness. You know, Quanta Services is another company that could fit this mold because, again, I just think this trend is so compelling. Over the next decade, right? And if you are looking over the long term, I think you do want some of these companies, even if they might underperform the next year or two as the focus really moves to behind the meter. So that's that's a very long way of saying I think American Superconductor, they missed on their earnings. Uh they were punished. It's a stock that's been somewhat of a roller coaster over the past two years. I don't think uh what they saw necessarily speaks to their future. I think you need to be prepared for volatility and lumpiness. And second, I think within the energy space, it's thinking about making sure your portfolio has companies engaged across the broader delivery in the space. And that is, again, why we're gonna have an expert from natural gas uh later this month. You know, we'll have the robotics expert, we'll have the robot, uh, the natural gas expert. And Austin, to the robotics point as well. Um, I did want to do more of a deep dive into robotics this week. We had some schedule changes and other things that we did more of a standard episode, but that is still coming and that still will happen in October. So I think that will answer the second part of this question in a lot more detail.

SPEAKER_00 45:07

Yes, I'm really looking forward to those interviews and that robotics deep dive. Apologies to our listeners. We just with the last couple of weeks we have had some moving pieces schedule-wise. So thank you for staying with us. Um, I love your commentary there to Dan on you know volatility and risk and the expectation that this is a roller coaster. And just would want to end on that note that volatility and risk are different, right? A roller coaster is really volatile. It's up, it's down, it's left, it's right, whatever, but it's not fundamentally risky, right? People, everyone's surveying roller coasters take you know thousands of riders per year safely. It's part of the ride, and investing in many ways can be the same way. Like, don't confuse volatility with risk. If you wanted the sort of returns that Netflix or Amazon or even NVIDIA put up, you had to sustain some pretty major drawdowns. And that is just part of it, is the price of emission. So, Dan, while American Superconductor has had disappointing guidance and last earnings would have been down, we do spend a lot of time commenting on earnings. Fundamentally, these are living, breathing companies that you want to bet on for the long run. Not all of them are going to work, but if you want the sort of stratospheric returns that Eric and other super investors have achieved over the years, you do have to look beyond individual earnings, even individual years and bet on a company for the long run, because again, volatility and risk are different. Volatility is things swinging all over the place, risk is permanent loss of capital, and um sometimes that volatility is the price of admission. So I don't know what the future is for American superconductor, but I would expect many more stocks in the portfolio to be on that roller coaster journey and a small number to become supernaturally successful, which carries your which carries your returns. Um, Eric, any other final notes for our listeners?

SPEAKER_01 46:43

No, I think I think that's it. As I noted, we we talked about last week. We we're gonna dive a lot deeper into robotics and and also natural gas. Those will be our two big themes for this month. So if you're there, if you're out there and you're looking for new investment ideas, they are coming. We're we're gonna have a lot this month. So um yeah, I I think that's that's kind of it for this week.

SPEAKER_00 47:04

In interviews, new investment ideas, it's gonna be a great month. Uh, Eric, I will leave you at that. Listeners, thank you so much for your time. We would appreciate your comments and feedback. If you can take a moment to subscribe and give us your thoughts on YouTube, Spotify, or wherever you listen to this podcast, we would appreciate it. Eric, thank you so much for your time. I'll see you next week. The AI Investor Podcast is for educational purposes only and should not be considered investment advice.